Performance Assessment of Control Loops

The design of advanced control algorithms has largely preoccupied control practitioners' efforts. The rationale has been that systems which are difficult to control need advanced optimal, non-linear, adaptive or similar control algorithms for better regulation. Although there are a variety of advanced control design techniques, few techniques exist for computing objective measures of control loop performance from routine operating industrial process data. The purpose of this book is to show the reader the most recent techniques for univariate and multivariate controller performance assessment. Specifically, this text ponders questions such as the following:

Is your controller healthy?

Is it doing its job well?

How can one obtain a non-invasive or a model-free assessment of controller performance from routing closed-loop data?

The theoretical results are supported by simulation, experimental (in-laboratory) and industrial evaluations.

A typical industrial process includes thousands of control loops. Instrumentation technicians generally maintain and service these loops, but rather infrequently. Routine maintenance of such loops at optimal settings can save a typical industrial complex hundreds of thousands of dollars a year. The development of quality measures of performance for such control loops in industrial settings is the subject matter of this book. This book ponders questions such as: Is your controller healthy? Is it doing its job well? How can one obtain a non-invasive or a model-free assessment of controller performance from routine closed-loop data? To answer these questions Performance Assessment of Control Loops considers:

• Theoretical aspects of control loop performance assessment for

- Univariate and

- Multivariate loops

• Evaluations of such assessment techniques via

- Simulations

- Experimental evaluations on computer-interfaced pilot scale processes

- Industrial applications